Roads of the Roman Empire

The Itiner-e Digital Atlas — Routes, Stations & the Ancient Network

All Roads Lead to Rome

At its peak, the Roman road network stretched over 250,000 miles — an engineered web of stone and gravel that stitched together an empire spanning from the rain-soaked hills of Britannia to the sun-scorched deserts of Mesopotamia. These were not mere footpaths: they were graded, drained, and surfaced with polygonal basalt blocks, designed to move legions at speed and hold together the most complex administrative state the ancient world had ever known. The viae publicae — the great trunk roads like the Via Appia, the Via Egnatia, and the Via Domitia — served as the arteries of empire, carrying soldiers, merchants, tax collectors, and couriers across landscapes that had never before seen a straight line.

The data in this analysis comes from the Itiner-e digital atlas of ancient Roman roads (Brughmans, de Soto, Pazout & Bjerregaard Vahlstrup, 2024), a collaborative project by Aarhus University that reconstructs the network from archaeological evidence, historical itineraries, and the Pleiades ancient places gazetteer. Licensed CC BY 4.0.

---

The Road Network

import sys, platform
sys.path.append("..")

import pandas as pd
import geopandas as gpd
import matplotlib.pyplot as plt
import json

from db_connection import query_to_geodataframe, query_to_dataframe
from map_builder import (
    create_base_map, add_geodataframe_layer, add_point_markers,
    finalize_map, build_photo_popup,
)

# Bounding box: West 5.74, South 36.71, East 29.04, North 48.47
BBOX = "ST_MakeEnvelope(5.743307, 36.706750, 29.040825, 48.474916, 4326)"

# Load road segments within bounding box (simplified for performance)
segments = query_to_geodataframe(f"""
    SELECT segment_id, name, road_type, segment_certainty,
           construction_period, itinerary, description,
           length_m, lower_date, upper_date, source_url,
           ST_Simplify(geometry, 0.005) AS geom
    FROM roman_road_segments
    WHERE ST_Intersects(geometry, {BBOX})
""", geom_col="geom", crs=4326)

# Load places within bounding box (major settlements, forts, bridges only)
places = query_to_geodataframe(f"""
    SELECT pleiades_id, name, place_type, start_year, end_year, url,
           geometry AS geom
    FROM roman_road_places
    WHERE ST_Intersects(geometry, {BBOX})
      AND place_type IN ('major-settlement', 'fort', 'bridge')
""", geom_col="geom", crs=4326)

# Load user-curated POIs
with open("roman-roads-poi.json", "r", encoding="utf-8") as f:
    poi_data = json.load(f)

print(f"{len(segments)} road segments loaded")
print(f"{len(places)} Pleiades places loaded")
print(f"{len(poi_data)} user POIs loaded")

5201 road segments loaded
1438 Pleiades places loaded
5 user POIs loaded

By Road Type

type_counts = (
    segments.groupby("road_type")
    .size()
    .reset_index(name="count")
    .sort_values("count", ascending=False)
)
type_counts

	road_type	count
3	Secondary Road	2613
0	Main Road	2016
1	River	286
2	Sea Lane	286

By Construction Period

period_counts = (
    segments.groupby("construction_period")
    .size()
    .reset_index(name="count")
    .sort_values("count", ascending=False)
)
period_counts

	construction_period	count
64	Manius Aquilius (129-126 BCE)	78
104	Trajan (98-117 CE)	75
46	Gnaeus Egnatius (146 BCE)	52
85	Septimius Severus Caracalla and Geta (201 CE)	45
73	Maximinus Thrax (235-238 CE)	44
...	...	...
31	Constantine the Great (326 CE)	1
48	Hadrian (115-116 CE)	1
86	Septimius Severus Caracalla and Geta (210 CE)	1
96	Tetrarchy (306 CE)	1
99	Titus Quintius Flaminius (123 BCE)	1

112 rows × 2 columns

---

Stationes and Mansiones

# Display places table
display_places = places[["name", "place_type", "start_year", "end_year"]].copy()
display_places.columns = ["Name", "Type", "From", "To"]
display_places.sort_values("Name").head(30)

	Name	Type	From	To
828	Sardis/Hyde?	major-settlement	-750.0	640.0
1072	"Ponte Sanguinario" at Spoletium	bridge	-30.0	640.0
116	"Trajan's Bridge" at Drobeta-Turnu Severin	bridge	-30.0	640.0
873	(H)Alesion Pedion	bridge	-330.0	300.0
897	(H)Enna	major-settlement	-550.0	640.0
158	(L)Ibida	fort	300.0	640.0
1435	(Mons) Brisiacus	fort	-30.0	640.0
963	*Agruvium	major-settlement	-330.0	640.0
582	*Aureus Mons	fort	-30.0	640.0
1301	*Aurinia/Saturnia	major-settlement	-750.0	640.0
1025	*Brigobannis	fort	-30.0	300.0
150	*Caput Bovis	fort	-30.0	640.0
120	*Diana	fort	-30.0	640.0
944	*Foetes	fort	300.0	640.0
643	*Malves(i)a	major-settlement	1700.0	2100.0
30	*Matar	major-settlement	-330.0	640.0
786	*Ponte Navata	fort	-30.0	640.0
575	*Una	fort	-30.0	640.0
1230	Abella	major-settlement	-330.0	640.0
1226	Abellinum	major-settlement	NaN	NaN
506	Abrittus	major-settlement	-330.0	640.0
728	Abrud	fort	-30.0	300.0
1228	Acerrae	major-settlement	-330.0	640.0
57	Acharnai (N)	major-settlement	-550.0	300.0
738	Acidava	fort	-30.0	300.0
984	Acumincum	fort	-30.0	640.0
617	Ad Aquas?	fort	-30.0	640.0
1027	Ad Fines	fort	-30.0	640.0
760	Ad Flexum	fort	-30.0	640.0
621	Ad Herculem	fort	-30.0	640.0

---

The Empire’s Road Map

import folium

# Base map centered on study area (positron for European geography)
m = create_base_map(center=[42.59, 17.39], zoom=5, tiles="positron")

# --- Road segments by type ---
road_colors = {
    "Main Road": {"color": "#f97316", "weight": 3, "opacity": 0.9},
    "Secondary Road": {"color": "#d97706", "weight": 2, "opacity": 0.8},
    "Sea Lane": {"color": "#64748b", "weight": 2, "opacity": 0.6, "dashArray": "8 4"},
    "River": {"color": "#0ea5e9", "weight": 2, "opacity": 0.7},
}

for road_type, style in road_colors.items():
    subset = segments[segments["road_type"] == road_type]
    if len(subset) > 0:
        m = add_geodataframe_layer(
            m, subset,
            name=road_type,
            tooltip_fields=["name", "road_type", "construction_period"],
            style=style,
            highlight={"weight": style.get("weight", 2) + 2, "opacity": 1},
        )

# Remaining road types (if any not in the dict above)
known_types = set(road_colors.keys())
other = segments[~segments["road_type"].isin(known_types)]
if len(other) > 0:
    m = add_geodataframe_layer(
        m, other,
        name="Other Roads",
        tooltip_fields=["name", "road_type", "construction_period"],
        style={"color": "#a855f7", "weight": 1.5, "opacity": 0.6},
    )

# --- Places as GeoJSON layer (major settlements, forts, bridges) ---
m = add_geodataframe_layer(
    m, places,
    name="Pleiades Places",
    tooltip_fields=["name", "place_type"],
    style={"color": "#f43f5e", "weight": 0.5, "fillOpacity": 0.4, "radius": 3},
    show=False,
)

# --- User-curated POIs with photo support ---
for poi in poi_data:
    popup_html = build_photo_popup(
        poi,
        photo_url=poi.get("photo_url") or None,
        photo_caption=poi.get("photo_caption") or None,
    )
    folium.Marker(
        location=[poi["lat"], poi["lon"]],
        popup=folium.Popup(popup_html, max_width=450),
        tooltip=f"<b>{poi['name']}</b>",
        icon=folium.Icon(color="red", icon="star", prefix="fa"),
    ).add_to(m)

m = finalize_map(m)
m

Make this Notebook Trusted to load map: File -> Trust Notebook

---

Points of Interest

User-curated points of interest with photos and descriptions. Add entries to roman-roads-poi.json to populate this section.

if poi_data:
    poi_df = pd.DataFrame(poi_data)
    display_poi = poi_df[["name", "category", "description"]].copy()
    display_poi.columns = ["Name", "Category", "Description"]
    display_poi.style.set_properties(
        subset=["Description"], **{"min-width": "350px", "white-space": "normal"}
    )
else:
    print("No POIs added yet — edit roman-roads-poi.json to get started.")

---

Construction Timeline

import plotly.express as px

# Filter to segments with date information
dated = segments.dropna(subset=["lower_date"]).copy()
dated = dated[dated["lower_date"] != 0]

if len(dated) > 0:
    # Aggregate by road_type and lower_date for a manageable chart
    date_summary = (
        dated.groupby(["road_type", "lower_date"])
        .size()
        .reset_index(name="segment_count")
    )

    fig = px.scatter(
        date_summary,
        x="lower_date",
        y="road_type",
        size="segment_count",
        color="road_type",
        title="Road Construction by Period",
        labels={"lower_date": "Date (BCE/CE)", "road_type": "Road Type",
                "segment_count": "Segments"},
        color_discrete_map={
            "Main Road": "#f97316",
            "Secondary Road": "#d97706",
            "Sea Lane": "#64748b",
            "River": "#0ea5e9",
        },
    )
    fig.update_layout(height=400, showlegend=True)
    fig.show()
else:
    print("No dated segments in the study area.")

---

Static Overview

import contextily as cx

# Reproject to Web Mercator for contextily basemap tiles
segments_3857 = segments.to_crs(epsg=3857)

fig, ax = plt.subplots(1, 1, figsize=(14, 10))

color_map = {
    "Main Road": ("#f97316", 0.8, 0.9),
    "Secondary Road": ("#d97706", 0.4, 0.8),
    "Sea Lane": ("#64748b", 0.4, 0.6),
    "River": ("#0ea5e9", 0.4, 0.7),
}

for road_type, (color, lw, alpha) in color_map.items():
    subset = segments_3857[segments_3857["road_type"] == road_type]
    if len(subset) > 0:
        subset.plot(ax=ax, color=color, linewidth=lw, label=road_type, alpha=alpha)

# Remaining types
other_3857 = segments_3857[~segments_3857["road_type"].isin(color_map.keys())]
if len(other_3857) > 0:
    other_3857.plot(ax=ax, color="#a855f7", linewidth=0.3, label="Other", alpha=0.5)

# Add hillshade/terrain basemap (Esri — no API key needed)
cx.add_basemap(ax, source=cx.providers.Esri.WorldShadedRelief, zoom=5)

ax.set_title("Roman Road Network — Study Area", fontsize=14)
ax.set_axis_off()
ax.legend(loc="lower left", fontsize=8, framealpha=0.9)
plt.tight_layout()
plt.show()

---

Data Source & Bibliography

Itiner-e Digital Atlas

Brughmans, T., de Soto, P., Pazout, A. and Bjerregaard Vahlstrup, P. (2024). Itiner-e: the digital atlas of ancient roads. Aarhus University. Licensed CC BY 4.0.

Website: https://www.itiner-e.org

Pleiades Gazetteer

Place data linked via the Pleiades ancient world gazetteer (https://pleiades.stoa.org), a community-built geographic information system for ancient places.

---

Technical Details

Environment

print("Python:", sys.version.split()[0])
print("Platform:", platform.platform())
print("GeoPandas:", gpd.__version__)
print("Pandas:", pd.__version__)

Python: 3.11.9
Platform: Windows-10-10.0.26100-SP0
GeoPandas: 1.1.2
Pandas: 2.2.2

Data Summary

print(f"Road segments (study area): {len(segments)}")
print(f"Pleiades places (study area): {len(places)}")
print(f"User POIs: {len(poi_data)}")
print(f"CRS: {segments.crs}")
print(f"\nBounding box: W 5.74, S 36.71, E 29.04, N 48.47")

Road segments (study area): 5201
Pleiades places (study area): 1438
User POIs: 5
CRS: EPSG:4326

Bounding box: W 5.74, S 36.71, E 29.04, N 48.47

Feature Breakdown

type_cert = (
    segments.groupby(["road_type", "segment_certainty"])
    .size()
    .reset_index(name="count")
    .sort_values("count", ascending=False)
)
type_cert

	road_type	segment_certainty	count
7	Secondary Road	Conjectured	2446
1	Main Road	Conjectured	1874
5	Sea Lane	Conjectured	286
3	River	Conjectured	221
0	Main Road	Certain	138
6	Secondary Road	Certain	107
4	River	Hypothetical	65
8	Secondary Road	Hypothetical	60
2	Main Road	Hypothetical	4

place_type_counts = (
    places.groupby("place_type")
    .size()
    .reset_index(name="count")
    .sort_values("count", ascending=False)
)
place_type_counts

	place_type	count
1	fort	576
2	major-settlement	571
0	bridge	291

Notes / Decision Log

• Data source: Itiner-e nightly bulk NDJSON export, loaded via roman_roads_harvester.py into PostGIS (everglades_gis database)
• Tables: roman_road_segments (LINESTRING, 4326), roman_road_places (POINT, 4326) — separate from Everglades tables
• Bounding box: Study area filtered to core Roman Empire (W 5.74, S 36.71, E 29.04, N 48.47)
• CRS: EPSG:4326 (WGS84) — native CRS of the Itiner-e data
• Photo popups: POI data stored in roman-roads-poi.json, supports external image URLs via <img> tags
• Next steps: Add POI photos, expand narrative sections, regional detail maps